Generic snippet support. Unlike previous versions of snippet
[idzebra-moved-to-github.git] / index / retrieve.c
index 4b162c8..d3ca5ad 100644 (file)
@@ -1,4 +1,4 @@
-/* $Id: retrieve.c,v 1.65 2007-02-28 18:43:06 adam Exp $
+/* $Id: retrieve.c,v 1.71 2007-08-21 11:06:47 adam Exp $
    Copyright (C) 1995-2007
    Index Data ApS
 
@@ -34,14 +34,15 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 
 #include "index.h"
 #include <yaz/diagbib1.h>
+#include <yaz/snprintf.h>
 #include <direntz.h>
-
+#include <yaz/oid_db.h>
 
 #define ZEBRA_XML_HEADER_STR "<record xmlns=\"http://www.indexdata.com/zebra/\""
 
 static int zebra_create_record_stream(ZebraHandle zh, 
-                               Record *rec,
-                               struct ZebraRecStream *stream)
+                                      Record *rec,
+                                      struct ZebraRecStream *stream)
 {
     RecordAttr *recordAttr = rec_init_attr(zh->reg->zei, *rec);
 
@@ -121,8 +122,8 @@ static int parse_zebra_elem(const char *elem,
 
 int zebra_special_sort_fetch(ZebraHandle zh, zint sysno, ODR odr,
                              const char *elemsetname,
-                             oid_value input_format,
-                             oid_value *output_format,
+                             const Odr_oid *input_format,
+                             const Odr_oid **output_format,
                              char **rec_bufp, int *rec_lenp)
 {
     const char *retrieval_index;
@@ -133,11 +134,12 @@ int zebra_special_sort_fetch(ZebraHandle zh, zint sysno, ODR odr,
     int ord;
 
     /* only accept XML and SUTRS requests */
-    if (input_format != VAL_TEXT_XML && input_format != VAL_SUTRS)
+    if (oid_oidcmp(input_format, yaz_oid_recsyn_xml) 
+        && oid_oidcmp(input_format, yaz_oid_recsyn_sutrs))
     {
         yaz_log(YLOG_WARN, "unsupported format for element set zebra::%s", 
                 elemsetname);
-        *output_format = VAL_NONE;
+        *output_format = 0;
         return YAZ_BIB1_NO_SYNTAXES_AVAILABLE_FOR_THIS_REQUEST;
     }
     
@@ -183,9 +185,9 @@ int zebra_special_sort_fetch(ZebraHandle zh, zint sysno, ODR odr,
         zebra_term_untrans(zh, index_type, dst_buf, str);
         
 
-        if (input_format == VAL_TEXT_XML)
+        if (!oid_oidcmp(input_format, yaz_oid_recsyn_xml))
         {
-            *output_format = VAL_TEXT_XML;
+            *output_format = yaz_oid_recsyn_xml;
             wrbuf_printf(wrbuf, ZEBRA_XML_HEADER_STR
                          " sysno=\"" ZINT_FORMAT "\""
                          " set=\"zebra::index%s/\">\n",
@@ -198,9 +200,9 @@ int zebra_special_sort_fetch(ZebraHandle zh, zint sysno, ODR odr,
             wrbuf_printf(wrbuf, "</index>\n");
             wrbuf_printf(wrbuf, "</record>\n");
         }
-        else if (input_format == VAL_SUTRS)
+        else if (!oid_oidcmp(input_format, yaz_oid_recsyn_sutrs))
         {
-            *output_format = VAL_SUTRS;
+            *output_format = yaz_oid_recsyn_sutrs;
             
             wrbuf_printf(wrbuf, "%s %c %s\n", string_index, index_type,
                          dst_buf);
@@ -208,7 +210,7 @@ int zebra_special_sort_fetch(ZebraHandle zh, zint sysno, ODR odr,
         *rec_lenp = wrbuf_len(wrbuf);
         *rec_bufp = odr_malloc(odr, *rec_lenp);
         memcpy(*rec_bufp, wrbuf_buf(wrbuf), *rec_lenp);
-        wrbuf_free(wrbuf, 1);
+        wrbuf_destroy(wrbuf);
         return 0;
     }
 }
@@ -216,26 +218,27 @@ int zebra_special_sort_fetch(ZebraHandle zh, zint sysno, ODR odr,
 int zebra_special_index_fetch(ZebraHandle zh, zint sysno, ODR odr,
                               Record rec,
                               const char *elemsetname,
-                              oid_value input_format,
-                              oid_value *output_format,
+                              const Odr_oid *input_format,
+                              const Odr_oid **output_format,
                               char **rec_bufp, int *rec_lenp)
 {
     const char *retrieval_index;
     size_t retrieval_index_len; 
     const char *retrieval_type;
     size_t retrieval_type_len;
-    WRBUF wrbuf = 0;
     zebra_rec_keys_t keys;
+    int ret_code = 0;
     
     /* set output variables before processing possible error states */
     /* *rec_lenp = 0; */
 
     /* only accept XML and SUTRS requests */
-    if (input_format != VAL_TEXT_XML && input_format != VAL_SUTRS)
+    if (oid_oidcmp(input_format, yaz_oid_recsyn_xml)
+        && oid_oidcmp(input_format, yaz_oid_recsyn_sutrs))
     {
         yaz_log(YLOG_WARN, "unsupported format for element set zebra::%s", 
                 elemsetname);
-        *output_format = VAL_NONE;
+        *output_format = 0;
         return YAZ_BIB1_NO_SYNTAXES_AVAILABLE_FOR_THIS_REQUEST;
     }
 
@@ -271,22 +274,27 @@ int zebra_special_index_fetch(ZebraHandle zh, zint sysno, ODR odr,
     zebra_rec_keys_set_buf(keys, rec->info[recInfo_delKeys],
                            rec->size[recInfo_delKeys], 0);
 
-    wrbuf = wrbuf_alloc();
-    if (zebra_rec_keys_rewind(keys)){
+    if (!zebra_rec_keys_rewind(keys))
+    {
+        ret_code = YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
+    }
+    else
+    {
         size_t slen;
         const char *str;
         struct it_key key_in;
-
-        if (input_format == VAL_TEXT_XML)
+        WRBUF wrbuf = wrbuf_alloc();
+    
+        if (!oid_oidcmp(input_format, yaz_oid_recsyn_xml))
         {
-            *output_format = VAL_TEXT_XML;
+            *output_format = input_format;
             wrbuf_printf(wrbuf, ZEBRA_XML_HEADER_STR
                          " sysno=\"" ZINT_FORMAT "\""
                          " set=\"zebra::index%s/\">\n",
                          sysno, elemsetname);
         }
-        else if (input_format == VAL_SUTRS)
-            *output_format = VAL_SUTRS;
+        else if (!oid_oidcmp(input_format, yaz_oid_recsyn_sutrs))
+            *output_format = input_format;
 
         while (zebra_rec_keys_read(keys, &str, &slen, &key_in))
         {
@@ -317,7 +325,8 @@ int zebra_special_index_fetch(ZebraHandle zh, zint sysno, ODR odr,
                     zebra_term_untrans(zh, index_type, dst_buf, str);
                     if (strlen(dst_buf))
                     {
-                        if (input_format == VAL_TEXT_XML){
+                        if (!oid_oidcmp(input_format, yaz_oid_recsyn_xml))
+                        {
                             wrbuf_printf(wrbuf, "  <index name=\"%s\"", 
                                          string_index);
                             
@@ -329,7 +338,8 @@ int zebra_special_index_fetch(ZebraHandle zh, zint sysno, ODR odr,
                             wrbuf_xmlputs(wrbuf, dst_buf);
                             wrbuf_printf(wrbuf, "</index>\n");
                         }
-                        else if (input_format == VAL_SUTRS){
+                        else 
+                        {
                             wrbuf_printf(wrbuf, "%s ", string_index);
                             
                             wrbuf_printf(wrbuf, "%c", index_type);
@@ -337,8 +347,8 @@ int zebra_special_index_fetch(ZebraHandle zh, zint sysno, ODR odr,
                             for (i = 1; i < key_in.len; i++)
                                 wrbuf_printf(wrbuf, " " ZINT_FORMAT, 
                                              key_in.mem[i]);
-
-                        /* zebra_term_untrans(zh, index_type, dst_buf, str); */
+                            
+                            /* zebra_term_untrans(zh, index_type, dst_buf, str); */
                             wrbuf_printf(wrbuf, " %s", dst_buf);
                         
                             wrbuf_printf(wrbuf, "\n");
@@ -348,15 +358,15 @@ int zebra_special_index_fetch(ZebraHandle zh, zint sysno, ODR odr,
                 }
             }
         }
-        if (input_format == VAL_TEXT_XML)
+        if (!oid_oidcmp(input_format, yaz_oid_recsyn_xml))
             wrbuf_printf(wrbuf, "</record>\n");
+        *rec_lenp = wrbuf_len(wrbuf);
+        *rec_bufp = odr_malloc(odr, *rec_lenp);
+        memcpy(*rec_bufp, wrbuf_buf(wrbuf), *rec_lenp);
+        wrbuf_destroy(wrbuf);
     }
-    *rec_lenp = wrbuf_len(wrbuf);
-    *rec_bufp = odr_malloc(odr, *rec_lenp);
-    memcpy(*rec_bufp, wrbuf_buf(wrbuf), *rec_lenp);
-    wrbuf_free(wrbuf, 1);
     zebra_rec_keys_close(keys);
-    return 0;
+    return ret_code;
 }
 
 
@@ -390,11 +400,135 @@ static void retrieve_puts_int(WRBUF wrbuf, const char *name,
     wrbuf_printf(wrbuf, "%s %i\n", name, value);
 }
 
-int zebra_special_fetch(ZebraHandle zh, zint sysno, int score, ODR odr,
-                           const char *elemsetname,
-                           oid_value input_format,
-                           oid_value *output_format,
-                           char **rec_bufp, int *rec_lenp)
+
+static void snippet_xml_record(ZebraHandle zh, WRBUF wrbuf, zebra_snippets *doc)
+{
+    const zebra_snippet_word *doc_w;
+    int mark_state = 0;
+
+    wrbuf_printf(wrbuf, "%s>\n", ZEBRA_XML_HEADER_STR);
+    for (doc_w = zebra_snippets_constlist(doc); doc_w; doc_w = doc_w->next)
+    {
+        if (doc_w->mark)
+        {
+            int index_type;
+            const char *db = 0;
+            const char *string_index = 0;
+
+            zebraExplain_lookup_ord(zh->reg->zei, doc_w->ord, 
+                                    &index_type, &db, &string_index);
+
+            if (mark_state == 0)
+            {
+                wrbuf_printf(wrbuf, "  <snippet name=\"%s\"",  string_index);
+                wrbuf_printf(wrbuf, " type=\"%c\">", index_type);
+            }
+            if (doc_w->match)
+                wrbuf_puts(wrbuf, "<s>");
+            /* not printing leading ws */
+            if (mark_state || !doc_w->ws || doc_w->match) 
+                wrbuf_xmlputs(wrbuf, doc_w->term);
+            if (doc_w->match)
+                wrbuf_puts(wrbuf, "</s>");
+        }
+        else if (mark_state == 1)
+        {
+            wrbuf_puts(wrbuf, "</snippet>\n");
+        }
+        mark_state = doc_w->mark;
+    }
+    if (mark_state == 1)
+    {
+        wrbuf_puts(wrbuf, "</snippet>\n");
+    }
+    wrbuf_printf(wrbuf, "</record>");
+}
+
+int zebra_special_snippet_fetch(ZebraHandle zh, const char *setname,
+                                zint sysno, ODR odr,
+                                const char *elemsetname,
+                                const Odr_oid *input_format,
+                                const Odr_oid **output_format,
+                                char **rec_bufp, int *rec_lenp)
+{
+    int return_code = 0;
+    Record rec;
+    
+    rec = rec_get(zh->reg->records, sysno);
+    if (!rec)
+    {
+        yaz_log(YLOG_WARN, "rec_get fail on sysno=" ZINT_FORMAT, sysno);
+        return_code = YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
+    }
+    else
+    {
+        const char *file_type = rec->info[recInfo_fileType];
+        void *recTypeClientData;
+        RecType rt = recType_byName(zh->reg->recTypes, zh->res,
+                                    file_type, &recTypeClientData);
+        zebra_snippets *hit_snippet = zebra_snippets_create();
+        WRBUF wrbuf = wrbuf_alloc();
+
+        zebra_snippets_hit_vector(zh, setname, sysno, hit_snippet);
+        
+        if (!rt)
+            return_code = YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
+        else
+        {
+            struct ZebraRecStream stream;
+            
+            return_code = zebra_create_record_stream(zh, &rec, &stream);
+            if (return_code == 0)
+            {
+                zebra_snippets *rec_snippet = zebra_snippets_create();
+                extract_snippet(zh, rec_snippet, &stream,
+                                rt, recTypeClientData);
+
+#if 0
+                /* for debugging purposes */
+                yaz_log(YLOG_LOG, "---------------------------");
+                yaz_log(YLOG_LOG, "REC SNIPPET:");
+                zebra_snippets_log(rec_snippet, YLOG_LOG, 1);
+                yaz_log(YLOG_LOG, "---------------------------");
+                yaz_log(YLOG_LOG, "HIT SNIPPET:");
+                zebra_snippets_log(hit_snippet, YLOG_LOG, 1);
+#endif
+
+                zebra_snippets_ring(rec_snippet, hit_snippet, 5, 5);
+
+#if 0
+                yaz_log(YLOG_LOG, "---------------------------");
+                yaz_log(YLOG_LOG, "RING SNIPPET:");
+                zebra_snippets_log(rec_snippet, YLOG_LOG, 1);
+#endif
+                
+                snippet_xml_record(zh, wrbuf, rec_snippet);
+
+                *output_format = yaz_oid_recsyn_xml;
+
+                
+                zebra_snippets_destroy(rec_snippet);
+            }
+            stream.destroy(&stream);
+        }
+        if (return_code == 0)
+        {
+            *rec_lenp = wrbuf_len(wrbuf);
+            *rec_bufp = odr_strdup(odr, wrbuf_cstr(wrbuf));
+        }
+        wrbuf_destroy(wrbuf);
+        rec_free(&rec);
+        zebra_snippets_destroy(hit_snippet);
+    }
+    return return_code;
+}
+
+int zebra_special_fetch(ZebraHandle zh, const char *setname,
+                        zint sysno, int score, ODR odr,
+                        const char *elemsetname,
+                        const Odr_oid *input_format,
+                        const Odr_oid **output_format,
+                        char **rec_bufp, int *rec_lenp)
 {
     Record rec;
     
@@ -402,30 +536,37 @@ int zebra_special_fetch(ZebraHandle zh, zint sysno, int score, ODR odr,
     /* *rec_lenp = 0; */
 
 
+    if (elemsetname && 0 == strcmp(elemsetname, "snippet"))
+    {
+        return zebra_special_snippet_fetch(zh, setname, sysno, odr,
+                                           elemsetname + 7,
+                                           input_format, output_format,
+                                           rec_bufp, rec_lenp);
+    }
 
     /* processing zebra::meta::sysno elemset without fetching binary data */
     if (elemsetname && 0 == strcmp(elemsetname, "meta::sysno"))
     {
         int ret = 0;
         WRBUF wrbuf = wrbuf_alloc();
-        if (input_format == VAL_SUTRS)
+        if (!oid_oidcmp(input_format, yaz_oid_recsyn_sutrs))
         {
             wrbuf_printf(wrbuf, ZINT_FORMAT, sysno);
-            *output_format = VAL_SUTRS;
+            *output_format = input_format;
         } 
-        else if (input_format == VAL_TEXT_XML)
+        else if (!oid_oidcmp(input_format, yaz_oid_recsyn_xml))
         {
             wrbuf_printf(wrbuf, ZEBRA_XML_HEADER_STR
                          " sysno=\"" ZINT_FORMAT "\"/>\n",
                          sysno);
-            *output_format = VAL_TEXT_XML;
+            *output_format = input_format;
         }
        *rec_lenp = wrbuf_len(wrbuf);
         if (*rec_lenp)
-            *rec_bufp = odr_strdup(odr, wrbuf_buf(wrbuf));
+            *rec_bufp = odr_strdup(odr, wrbuf_cstr(wrbuf));
         else
             ret = YAZ_BIB1_NO_SYNTAXES_AVAILABLE_FOR_THIS_REQUEST;
-        wrbuf_free(wrbuf, 1);
+        wrbuf_destroy(wrbuf);
         return ret;
     }
 
@@ -466,7 +607,8 @@ int zebra_special_fetch(ZebraHandle zh, zint sysno, int score, ODR odr,
     }
 
     /* only accept XML and SUTRS requests from now */
-    if (input_format != VAL_TEXT_XML && input_format != VAL_SUTRS)
+    if (oid_oidcmp(input_format, yaz_oid_recsyn_xml)
+        && oid_oidcmp(input_format, yaz_oid_recsyn_sutrs))
     {
         yaz_log(YLOG_WARN, "unsupported format for element set zebra::%s", 
                 elemsetname);
@@ -481,9 +623,9 @@ int zebra_special_fetch(ZebraHandle zh, zint sysno, int score, ODR odr,
         WRBUF wrbuf = wrbuf_alloc();
         RecordAttr *recordAttr = rec_init_attr(zh->reg->zei, rec); 
 
-        if (input_format == VAL_TEXT_XML)
+        if (!oid_oidcmp(input_format, yaz_oid_recsyn_xml))
         {
-            *output_format = VAL_TEXT_XML;
+            *output_format = input_format;
             
             wrbuf_printf(wrbuf, ZEBRA_XML_HEADER_STR
                          " sysno=\"" ZINT_FORMAT "\"", sysno);
@@ -501,9 +643,9 @@ int zebra_special_fetch(ZebraHandle zh, zint sysno, int score, ODR odr,
                          recordAttr->recordSize,
                          elemsetname);
         }
-        else if (input_format == VAL_SUTRS)
+        else if (!oid_oidcmp(input_format, yaz_oid_recsyn_sutrs))
         {
-            *output_format = VAL_SUTRS;
+            *output_format = input_format;
             wrbuf_printf(wrbuf, "sysno " ZINT_FORMAT "\n", sysno);
             retrieve_puts_str(wrbuf, "base", rec->info[recInfo_databaseName]);
             retrieve_puts_str(wrbuf, "file", rec->info[recInfo_filename]);
@@ -521,11 +663,11 @@ int zebra_special_fetch(ZebraHandle zh, zint sysno, int score, ODR odr,
         }
        *rec_lenp = wrbuf_len(wrbuf);
         if (*rec_lenp)
-            *rec_bufp = odr_strdup(odr, wrbuf_buf(wrbuf));
+            *rec_bufp = odr_strdup(odr, wrbuf_cstr(wrbuf));
         else
             ret = YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
 
-        wrbuf_free(wrbuf, 1);
+        wrbuf_destroy(wrbuf);
         rec_free(&rec);
         return ret;
     }
@@ -548,10 +690,11 @@ int zebra_special_fetch(ZebraHandle zh, zint sysno, int score, ODR odr,
 }
 
                           
-int zebra_record_fetch(ZebraHandle zh, zint sysno, int score,
+int zebra_record_fetch(ZebraHandle zh, const char *setname,
+                       zint sysno, int score,
                        zebra_snippets *hit_snippet, ODR odr,
-                       oid_value input_format, Z_RecordComposition *comp,
-                       oid_value *output_format,
+                       const Odr_oid *input_format, Z_RecordComposition *comp,
+                       const Odr_oid **output_format,
                        char **rec_bufp, int *rec_lenp, char **basenamep,
                        char **addinfo)
 {
@@ -569,7 +712,7 @@ int zebra_record_fetch(ZebraHandle zh, zint sysno, int score,
 
     /* processing zebra special elementset names of form 'zebra:: */
     if (elemsetname && 0 == strncmp(elemsetname, "zebra::", 7))
-        return  zebra_special_fetch(zh, sysno, score, odr,
+        return  zebra_special_fetch(zh, setname, sysno, score, odr,
                                     elemsetname + 7,
                                     input_format, output_format,
                                     rec_bufp, rec_lenp);
@@ -600,7 +743,6 @@ int zebra_record_fetch(ZebraHandle zh, zint sysno, int score,
 
     if (rec)
     {
-       zebra_snippets *snippet;
        zebra_rec_keys_t reckeys = zebra_rec_keys_open();
         RecType rt;
         struct recRetrieveCtrl retrieveCtrl;
@@ -631,22 +773,6 @@ int zebra_record_fetch(ZebraHandle zh, zint sysno, int score,
        zebra_rec_keys_to_snippets(zh, reckeys, retrieveCtrl.doc_snippet);
        zebra_rec_keys_close(reckeys);
 
-#if 0
-       /* for debugging purposes */
-       yaz_log(YLOG_LOG, "DOC SNIPPET:");
-       zebra_snippets_log(retrieveCtrl.doc_snippet, YLOG_LOG);
-       yaz_log(YLOG_LOG, "HIT SNIPPET:");
-       zebra_snippets_log(retrieveCtrl.hit_snippet, YLOG_LOG);
-#endif
-       snippet = zebra_snippets_window(retrieveCtrl.doc_snippet,
-                                       retrieveCtrl.hit_snippet,
-                                       10);
-#if 0
-       /* for debugging purposes */
-       yaz_log(YLOG_LOG, "WINDOW SNIPPET:");
-       zebra_snippets_log(snippet, YLOG_LOG);
-#endif
-
         if (!(rt = recType_byName(zh->reg->recTypes, zh->res,
                                   file_type, &clientData)))
         {
@@ -669,7 +795,6 @@ int zebra_record_fetch(ZebraHandle zh, zint sysno, int score,
             *addinfo = retrieveCtrl.addinfo;
         }
 
-       zebra_snippets_destroy(snippet);
         zebra_snippets_destroy(retrieveCtrl.doc_snippet);
 
         stream.destroy(&stream);